Government to Promote Development of Korean-style ChatGPT

Launch of Three Sector Working Groups... Exploring New System Directions
Focusing on Developing High-Quality Corpora Reflecting Korean Language Characteristics

The Ministry of Culture, Sports and Tourism announced on the 22nd that it will launch three working groups to prepare for the changes brought by generative AI technologies such as ChatGPT. The three areas are AI copyright law system improvement, Korean corpus training for 'AI proficient in Korean,' and AI utilization in content creation and industry. The ministry will work closely with experts to find policy solutions. Minister Park Bo-gyun stated, "We will support all companies to learn from abundant data without worrying about copyright issues and data collection costs."

Minister of Culture, Sports and Tourism Park Bo-gyun is attending the National Assembly's audit of the Ministry of Culture, Sports and Tourism held on the 5th, responding to lawmakers' questions. Photo by Yoon Dong-joo doso7@

The AI Copyright Law System Improvement Working Group will take its first step on the 24th at the Seoul office of the Korea Copyright Commission. It will discuss smooth usage methods for copyrighted works used as AI training data, the legal status of AI outputs and their recognition under copyright law, copyright infringement issues arising from AI technology use, and responsibility regulations. Participants include copyright scholars, legal professionals, AI industry representatives, and creators, seeking a system suitable for the changing era. A Ministry of Culture, Sports and Tourism official said, "We will find a consensus that supports AI development as a new industry, such as preparing 'Guidelines for AI Output Utilization from a Copyright Perspective,' while fairly guaranteeing the rights of human creators."

The Korean Corpus Training Working Group focuses on developing high-quality corpora reflecting the characteristics of the Korean language. Most generative AI, including ChatGPT, was developed in English-speaking countries and has not sufficiently learned Korean. The ministry will enhance the quality of thirty-seven types of Korean corpora (about 2.2 billion morphemes) built since 2018 to fill the gaps. The task force includes corpus construction and academic experts, as well as researchers from the National Institute of the Korean Language. They will identify the demand for corpora necessary for AI technology development and create 1 billion morphemes of high-quality corpora reflecting Korean characteristics by 2027. A ministry official said, "To accelerate the development of Korean-style ChatGPT, we will distribute twenty-five types (about 120 million morphemes) of Korean corpora this year," adding, "We will pilot an evaluation system to verify whether AI language models understand and generate Korean well and possess knowledge of Korean socio-cultural contexts."

The AI Utilization in Content Creation and Industry Working Group is effectively the AI task force for the content sector. It consists of nine members, including Kim Jin-hyung, a pioneer of AI in Korea and emeritus professor at KAIST, along with related academic and industry stakeholders. At the first meeting on the 15th, they decided to concretize plans for creating a trustworthy content production environment using AI, spreading AI utilization mindsets among content planners and producers, fostering AI content startups for private-led market growth, and collaborating with large corporations. A ministry official said, "We will focus on identifying the impact of generative AI on the content industry and various issues, investigating and analyzing cases of AI application in the content industry, and proposing related policy support measures."

This content was produced with the assistance of AI translation services.